At Prevedere, I architected an operational forecasting platform for retail supply chain and manufacturing inventory data, helping enterprises generate demand and inventory forecasts.
I built an ML forecasting pipeline that orchestrates Prophet and LightGBM models with Kubernetes jobs, and added a browser-based analytics layer using DuckDB-WASM and Angular.
I also led the migration from SAS to a Python/R-based architecture, completing it six months ahead of schedule and reducing infrastructure costs. By redesigning data retrieval and caching, I cut API response times from 2 seconds to 50ms.
Earlier at Prevedere and OEConnection, I developed predictive modeling capabilities, pricing APIs, and automated tests. I’ve also served as lead engineer across feature teams and mentored engineers.

